FedRFQ: Prototype-Based Federated Learning with Reduced Redundancy, Minimal Failure, and Enhanced Quality

TL;DR

FedRFQ is a novel federated learning method that reduces prototype redundancy, enhances robustness against failures and attacks, and improves accuracy on non-IID data through innovative mechanisms.

Contribution

The paper introduces FedRFQ, a prototype-based federated learning approach with SoftPool and BFT-detect to address redundancy, failures, and security issues, advancing non-IID data handling.

Findings

FedRFQ outperforms baselines in accuracy on MNIST, FEMNIST, and CIFAR-10.

SoftPool effectively mitigates prototype redundancy and failure.

BFT-detect enhances security against poisoning and server malfunctions.

Abstract

Federated learning is a powerful technique that enables collaborative learning among different clients. Prototype-based federated learning is a specific approach that improves the performance of local models under non-IID (non-Independently and Identically Distributed) settings by integrating class prototypes. However, prototype-based federated learning faces several challenges, such as prototype redundancy and prototype failure, which limit its accuracy. It is also susceptible to poisoning attacks and server malfunctions, which can degrade the prototype quality. To address these issues, we propose FedRFQ, a prototype-based federated learning approach that aims to reduce redundancy, minimize failures, and improve \underline{q}uality. FedRFQ leverages a SoftPool mechanism, which effectively mitigates prototype redundancy and prototype failure on non-IID data. Furthermore, we introduce the BFT-detect, a BFT (Byzantine Fault Tolerance) detectable aggregation algorithm, to ensure the security of FedRFQ against poisoning attacks and server malfunctions. Finally, we conduct experiments on three different datasets, namely MNIST, FEMNIST, and CIFAR-10, and the results demonstrate that FedRFQ outperforms existing baselines in terms of accuracy when handling non-IID data.